Proc. Natl. Acad. Sci. USAVol. 90, pp. 562-566, January 1993Medical Sciences

Changes in drug sensitivity of human immunodeficiency virus type1 during therapy with azidothymidine, dideoxycytidine, anddideoxyinosine: An in vitro comparative studyTAKUMA SHIRASAKA*, ROBERT YARCHOANt, MARY C. O'BRIEN*, ROBERT N. HUSSON*,BARRY D. ANDERSON*, Eui KoJIMA*, TAKASHI SHIMADAt, SAMUEL BRODER§, AND HIROAKI MITSUYA*¶*Experimental Retrovirology Section, tRetrovirology Diseases Section, Medicine Branch, and Clinical Hematology Branch, National Heart, Lung and BloodInstitute, National Cancer Institute, Bethesda, MD 20892

Communicated by Howard M. Temin, September 11, 1992 (received for review April 22, 1992)

ABSTRACT Human immunodeficiency virus type 1(1IV-1) strains were Isolated from nine patients before andafter prolonged therapy with either an alternating regimen of3'-azido-3'-deoxythymidine (AZT) and 2',3'-dldeoxycytldin(ddC) (AZT/ddC) or 2',3'-deoxyinosine (ddl) alone. ANstrains obtained from four patients who received AZT/ddC forup to 41 mo were highly insensitive to AZT in vitro. Only onestrain obtained after AZT/ddC therapy showed reduced sus-ceptibility to ddC in addition to AZT and had previouslyunreported amino acid substitutions in the viral polymerase-ending pol region, whereas three other strains had one ormore of the five previously reported AZT-related mutations. Inflve HIV-1 strains from patents who received ddI for up to 29mo, no appreciable decrease in sensitivity to ddl was detect.Two strains Isolated after ddI therapy had no t aminoacid mutations, although three strains had a mutation report-edly acited with ddI administration. These data suggestthat HIV-1 develops reduced susceptibility to AZT morereadily than to ddC and ddI and/or that the reduced suscep-tibility to ddC and ddI is modest in degree. Moreover, thepresent data suggest that an alternating regimen of AZT andddC does not block the emergence ofAZT- ensitve va .It should be noted, however, that the current results do notprovide a basds for concluding that AZT/ddC or ddI is inferior,equivalent, or superior to AZT as therapy of AIDS.

The ability to provide effective long-term antiretroviral ther-apy using single agents for human immunodeficiency virustype 1 (HIV-1) infection became a complex issue when HIV-1strains that were less susceptible to 3'-azido-3'-deoxythymi-dine (AZT or zidovudine) in vitro were isolated from patientswith AIDS who received long-term AZT therapy (1-3). Two2',3'-dideoxynucleosides, 2',3'-dideoxyinosine (ddN ordidanosine) and 2',3'-dideoxycytidine (ddC), have beenshown to have clinical activity in patients with HIV-1 infec-tion (4-8). St. Clair et al. (9) have recently reported thatpatients who received long-term AZT therapy and then ddltherapy developed reduced susceptibility to ddI and that theemergence of ddI insensitivity was associated with a rever-sion to a more AZT-sensitive phenotype. However, thebehavior of HIV-1 at the genetic and phenotypic level uponexposure to multiple antiretroviral agents is as yet poorlyunderstood. In the current study, we specifically asked howeasily HIV-1 develops reduced sensitivity to three drugs:AZT, ddC, and ddI, when given as single drugs or incombination. We also asked whether the addition of ddC toAZT therapy in an alternating regimen could block theemergence of AZT-insensitive HIV-1 variants.

MATERIALS AND METHODSNucleosides. ddC and ddl for in vitro use were provided by

the Developmental Therapeutics Program, Division of Can-cer Treatment, National Cancer Institute, whereas AZT waspurchased from Sigma.

of Clinical HIV-1 Strains. HIV-1 strains wereisolated from four patients with AIDS or AIDS-related com-plex (ARC) (patients 100, 101, 103, and 104) who received analternating regimen of AZT (7 days) and ddC (7 days)(AZT/ddC) for 15, 22, 41, and 32 mo, respectively. Patient103 received ddI for the last 3 mo in addition to AZT/ddCtherapy. HIV-1 strains were also isolated from five patientswith AIDS orARC (patients 200, 201, 203, 204, and 205), whoreceived ddl for 25, 20, 29, 18, and 20 mo, respectively.Patients 203, 204, and 205 had received AZT before ddl

therapy for 12, 3, and 12 mo, respectively.To isolate primary HIV-1 strains, peripheral blood mono-

nuclear cells (PBMCs) (1-2 x 106) from each patient werecocultured with phytohemagglutinin (PHA)-activated PB-MCs (PHA-PBMCs, 1-2 x 106) from HIV-1-seronegativevolunteers in 24-well microtiter culture plates in 2 ml ofculture medium [RPMI 1640 medium/15% heat-inactivatedfetal calf serum/recombinant interleukin 2 at 10 units per ml(Amgen Biologicals)/4 mM L-glutamine/penicillin at 50 unitsper ml/streptomycin at 50 ug/ml]. When the level ofp24 Gagprotein in the culture supernatant reached .10 ng/ml asassessed by RIA (Dupont), the supernatant was collected asa source of infectious virions.Determination of Sensitivity of Clinical HIV-1 Strains

Agant Drus. In the current study, each pair of pre- andposttherapy isolates to be compared was simultaneouslytitrated for infectivity and assessed for drug sensitivity. Todetermine changes in drug sensitivity, PHA-PBMCs werefirst exposed to various concentrations of drug for 2 hr, thenexposed to a 20x tissue culture 50% infectious dose (TCID50)of each virus isolate. To confirm that the viral dose used wasthe desired one, the inoculum was retitrated for TCID50 in thesame assay. The drug concentrations used were 0.0005,0.005, 0.05, 0.5, 5, 50 ,uM for AZT; 0.005, 0.05, 0.5, 2, 5 ,uMfor ddC; and 0.1, 1, 10, 50, 100 juM for ddM. The sensitivityof a given HIV-1 strain against a drug was defined as the drugconcentration that yielded 50% p24 Gag protein-negativewells (CN5o) on days 8-10 in culture. When concentrations ofp24 Gag protein were <0.6 ng/ml, representing the minimum

Abbreviations: AZT, 3'-azido-3'-deoxythymidine; CN5o, concentra-tion yielding 50%O negative wells; ddC, 2',3'-dideoxycytidine; ddI,2',3'-dideoxyinosine; HIV-1, human immunodeficiency virus type 1;PBMC, peripheral blood mononuclear cell; PHA, phytohemaggluti-nin; PHA-PBMC, PHA-activated PBMC; RT, reverse transcriptase;TCID5o, tissue culture 50%6 infectious dose.ITo whom reprint requests should be addressed at: ExperimentalRetrovirology Section, Medicine Branch, National Cancer Insti-tute, Building 10, Room SA11, Bethesda, MD 20892.

562

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement"in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Dow

nloa

ded

by g

uest

on

June

6, 2

020

Proc. Natl. Acad. Sci. USA 90 (1993) 563

detection limit (10), cultures were defined as negative wells.All culture assays were done in quadruplicate in 24-wellmicrotiter culture plates. The variability in CN50 values wasgenerally within a factor of -6 for AZT, -2 for ddC, and -2for ddM.

Determination of Nucleotide Sequences. Cell lysates fromcoculture of PHA-PBMCs with patient's PBMCs were sub-jected to PCR as previously described by Saiki et al. (11) withminor modifications. The primers used for amplification of a1.7-kb segment ofthepol gene were 5'-TTGCACTTTGAAT-TCTCCCATTAG-3' and 5'-CTTATCTATTCCATCTA-GAAATAGT-3', as described by Larder and Kemp (2). AfterPCR, the pol fragment amplified was digested with EcoRI andXba I and inserted into the pTZ19R vector. CompetentEscherichia coli (strain DH5a) were then transformed withthe pol-inserted pTZ19R, which was subsequently purifiedand sequenced by the dideoxynucleotide chain-terminationmethod.

En

LU

z

9_

Z o

Z

z0

a:

z(11z0u~

A B

0 1 2 3 0 1 2 3DURATION OF THERAPY (

CTo ddl

204V203

200

201205

0 1 2 3yr)

RESULTS

In Vitro Drug Sensitivity of HIV-1 Strains Against AZT,ddC, or ddI. HIV-1 strains were isolated from nine patientsbefore and after prolonged therapy with either an alternatingregimen ofAZT and ddC (AZT/ddC) or ddI alone. When weexamined four pairs of pre- and posttherapy HIV-1 strainsisolated from patients who had received AZT/ddC, 60- to200-fold differences in CN50 values for AZT were observed(Fig. 1). The sensitivity of HIV-11oopost to ddC and ddI,however, was essentially the same as that of HIV-1100pre (Fig.1 B and C). Similar results were obtained for HIV-1 strainsisolated from patients 101 and 104. In contrast, HIV-11o3pot41mo, isolated from patient 103 who received AZT/ddC for 41mo, appeared to have developed resistance to AZT, ddC, andddI, as discussed below.Five patients who received ddI were also examined. Three

of five patients had received AZT before ddI therapy andshowed a high level of insensitivity to AZT at entry (Fig. 2A,at time 0). After long-term ddI therapy, two patients' HIV-1strains (patient 203 and 204) showed a decrease in sensitivityto ddC by 2-fold and 7-fold, respectively (Fig. 2B). Bothpatients had never received ddC, and whether this level of

Un

LJ

z

0R

Z

XZ

UJ U

z0

z

z0CD

100 ATo AZT

B C

To ddl

103

104

101100

0 1 2 3 4

DURATION OF THERAPY (yr)

FIG. 1. In vitro drug sensitivity of pre- and posttherapy HIV-1strains isolated from patients who received an alternating regimen ofAZT (1 week) and ddC (1 week). Viral titration was done in eightreplicates, whereas determination of drug sensitivity was done inquadruplicate. Amino acid sequences of posttherapy strains frompatients 103 and 104 shown in Fig. 3 represent those of the strainsisolated at the 41st and 32nd mo of AZT/ddC therapy (each shownby an arrow in A), respectively.

FIG. 2. In vitro drug sensitivity of pre- and posttherapy HIV-1strains isolated from patients who received ddl. The posttherapystrain of patient 203 (HIV-1203post), the amino acid sequence ofwhichis shown in Fig. 4, was isolated at the 26th mo of ddI therapy (shownby a small arrow in A). The HIV-1 strain from patient 203 2 mo afterresumption of AZT therapy is shown by a large arrow in A.

phenotypic changes truly represents reduced sensitivity toddC is unclear. However, no detectable difference in thesensitivity to ddI between pre- and posttherapy HIV-1 strainsisolated from any of the five patients was detected (Fig. 2C).

Nucleotide and Amino Acid Sequences of Pre- and Post-AZT/ddC Strains. HIV-11oo0pog had an Asp-67 -- Asn sub-stitution in eight of eight clones sequenced [designated asAsp-67 -- Asn(8/8)] as well as three other previously re-ported substitutions [Lys-70 -+ Arg(8/8), Thr-215 -* Tyr(6/8), and Lys-219 -* Gln(8/8)] (Fig. 3). In HIV-11o4pst, twopatterns of substitutions, Thr-215 -k Tyr(6/8) and Thr-215 --

Asp(2/8), were identified at amino acid residue 215. Thelatter of these has not been previously described. A Val-106-- Ile(8/8) substitution was seen in HIV-11o4post. The signif-icance of this substitution is unclear.

In addition to the four known AZT-related substitutions,Asp-67-) Asn(8/8), Lys-70--+ Arg(8/8), Thr-215--* Phe(8/8),and Lys-219 -+ Gln(8/8), HIV-11olpost had a Thr-69 --

Asp(8/8) substitution. In this regard, Fitzgibbon et al. (12)have recently reported a Thr-69 -* Asp substitution in aHIV-1 strain isolated from a patient receiving ddC therapy.In our study, however, the Asp-69-carrying HIV-11olpog,showed no detectable decrease in sensitivity to ddC (Fig.1B), and three other HIV-1 strains isolated from patientsreceiving AZT/ddC did not have this substitution (Fig. 3).An AZT-Insensitive HIV-1 Strain That Has None of the

Previously Reported AZT-Related Amino Acid Mutations.HIV-1lO3post4lmo showed a substantial decrease in sensitivityto ddC (40-fold increase in CN50) and a relative insensitivityto ddI (4.5-fold increase in CN50) in addition to a high levelofAZT insensitivity (200-fold increase in CN50). Although anHIV-1 strain isolated from this patient after 6 mo ofAZT/ddCtherapy (HIV-1103postgmo) did not have significant amino acidsubstitutions, HIV-11o3post-41mo had seven substitutions,Ala-62 -- Val(8/8), Val-75 -- Ile(8/8), Phe-77 -* Leu(8/8),Phe-116 -* Tyr(8/8), Val-118 -+ Ile(2/8), Gln-151 -* Met(8/8), and Ile-202 -* Val(7/7), none of which has been describedpreviously, and lacked any of the previously reported AZT-related amino acid substitutions.

Nucleotide and Amino Acid Sequences of Pre- and Post-ddIHIV-1 Strains. Three of the five post-ddI isolates examined,HIV-12ooposts, HIV-12o1post, and HIV-1203post, were found tohave a Leu-74 -+ Val(8/8) substitution (Fig. 4). Patient 203

Medical Sciences: Shirasaka et al.

Dow

nloa

ded

by g

uest

on

June

6, 2

020

564 Medical Sciences: Shirasaka et al. Proc. Natl. Acad. Sci. USA 90 (1993)

VIRUSSTRAIN AMINO ACID SEQUENCE

30 40 50 60 70 80 90VKQWPLTEEK IKaLvEICTE MEKEGKISKi GPENPYNTPV FAIKKKDSTK WRKLVDFREL NkrTQDFWEV

iERS 10 Opr - ---------- -E ------- --- K------ ---------- ---------- ---------- ----------

ERS10 Opo0t ---------- ---------- ---------- ---------- -------N--R ---------- ----------

ERSl0 lpreERS10 1post - -- ------------------- ----- N-DR ---------- --------- IERS103pre -----------R---- ---------- ---------- ---------- ----------

ERSl03post-6*-o-- ---------- -G---- --- - ----------

ERS103post-410o ---------- ---------- ---------- ---------- - V-------- ----I-L--- ----------

ERS104pre -R -------- ---------K ---------- ---------- ------G--- ---------- ----------

ERS104post

100 110 120 130 140 150 160QLGIPHPAGL KkKKSVTVLD VGDAYFSVPL DkdFRKYTAF TIPSiNNETP gIRYQYNVLP QGWKGSPAIF

e

ERSlOOpr-Te- --- T ----- -- - --

ERS 10 Opost --- -- -- - -- - -- -- -- - -- - -- -- -- - -- - -- -- -- - -- - ----T -- --- - -- -- -- - -- - -- -- -- -- -

ERS10p-r.-T---------- ---------- ---------- ---------- ----T ----- ---------- ------L---ERS 101po-t---------- ---------- ---------- ---------- ----T----- -------------E -----

ERSlO03preERS103post-6mo ---------- ---------- ---------- N-N ------- ---------- -----H---- ----------

ERS103post-41mo ---------- ---------- -----Y-I-- ---------- ---------- ---------- M---------ERSlO4preERS104post ---------- ----- I ---- ---------- ---------- ---------- ---------- ----------

170 180 190 200 210 220 230QsSMTKILEP FrKQNPdIvI YQYMDDLYVG SDLEIgQHRt KIEELRqHLL rWGfT TPDKK HQKEPPFLWM

i e ka

ERS 10l Op reERS10 Opot ---------- ---------- ---------- -------- K- ---------- - ---Y ---Q----------

ERS 10 lpre ---------- ---------- ---------- ---------- IV -------- ----- ----- --------

ERS10lpo-t---- -- -------- - ---------- -V-------- ----F ---Q- ---------

ERS10 3pre ----- R ---- -- - ---------- --- - --------- ----- --- ----------

ERS103po0t-63o - R----R -----------------------E ---------- ---------- ----- ----- - -P--- ----

ERS103post-41mo-- -R---- ---------- ---------- -------- - V-------- ----- ----- ---------

ERS10 4pr,ERS104po t Y *

FIG. 3. Deduced amino acid sequences of reverse transcriptase (RT) in HIV-1 strains isolated from patients receiving AZT/ddC. The numberof clones sequenced was four for ERS1OOpr,, eight for ERS100pt, four for ERS101pr, eight for ERS101pos, four for ERS103pr., four forERS103post-6mo, eight for ERS103post41,mo, four for ERS104prc, and eight for ERS104p.,t. Letters for each isolate represent amino acids thatdiffered from the consensus sequence 1 of HIV-1 polymerase (19), shown at top as reference in >15% clones sequenced. Dashes indicate aminoacid residues conserved with respect to the reference sequence. In strain ERS100pre K -- E substitution at codon 32 was seen in four of four[designated as K32E(4/4)], E44K(4/4), and i135T(4/4). The following other substitutions were found: in strain ERS100post, D67N(8/8), K7°R(8/8),i'35T(8/8), R199K(8/8), T215Y(6/8), and K219Q(8/8); in strain ERS101pr,, i135T(4/4), P157L(1/4), K201I(1/4), and 1202V(3/4); in strain ERS101pos,D67N(8/8), T*9D(8/8), K70R(8/8), V9°I(8/8), i135T(8/8), K"54E(2/8), I202V(8/8), T215F(8/8), and K219Q(8/8); in strain ERS103pre, K46R(1/4) andK166R(4/4); in strain ERS103psto, E44G(1/4), D 21N(1/4), D123N(1/4), Y146H(1/4), K'66R(4/4), G'90E(1/4), and Q222P(1/4); in strainERSl03post-lmo, A62V(8/8), V75I(8/8), F77L(8/8), F'16Y(8/8), V118I(2/8), Q151M(8/8), K166R(7/7), and I202V(7/7); in strain ERS104pr,,K22R(1/4), E40K(1/4), and D67G(1/4); and in strain ERS104post, V106I(8/8), T215Y(6/8), and T215D(2/8, shown by *) were identified. Single-letteramino acid code is used. Amino acids shown in lowercase letters represent variable amino acid residues (19).

had received AZT therapy before ddI therapy and thenreceived ddI monotherapy. Patients 200 and 201, however,had received no prior AZT therapy. Therefore, these dataindicate that patients who received only ddI monotherapydeveloped the Leu-74 -+ Val substitution.

Restoration of Sensitivity to AZT of AZT-Insensitive HIV-1Strain after Chaning to ddI Therapy. Three patients, 203,204, and 205, had received AZT prior to ddI therapy. BeforeddI, each of these three patients' HIV-1 isolates showed asubstantial'reduced susceptibility to AZT (Fig. 2A at time 0).Interestingly, these strains resumed relative sensitivity toAZT after switching to ddM. HIV-12o4pre had the Lys-70Arg(3/4) substitution that should be related to the prior AZTtherapy; however, after 18 mo of ddI therapy followed by 3mo ofAZT therapy, this patient's posttherapy strain lost theLys-70 -* Arg substitution, reverting to the wild-type aminoacid Lys-70, while it acquired the Thr-215 -* Tyr substitutionthat is thought to correlate with insensitivity to AZT (2). Infact, the CN50 ofHIV-1204pot (16 1.M) was still relatively high,indicating that HIV-12wpost was yet substantially insensitiveto AZT (Fig. 2A).

HIV-1205' also had the AZT-related Lys-70 -* Arg(1/4)substitution at entry that was presumably from prior AZTtherapy. However, this substitution was not seen after 20 moof ddI therapy. Thus, the HIV-1 strain from patient 205 mayhave regained sensitivity to AZT by losing this substitutionafter ddI therapy.

Patient 203, who had previously received AZT for 12 mo,had a fall in his CD4 count after 26 mo of ddI therapy (Fig.2A), and he expressed a desire to try AZT. During the 26 moof ddI therapy, his HIV-1 had become relatively sensitive toAZT. Genetically the virus maintained the Thr-215 -* Tyrsubstitution, while acquiring a new Leu-74 -* Val substitu-tion (Table 1). This result suggests that the acquisition ofLeu-74 -- Val rendered the AZT-insensitive variant some-what sensitive to AZT, as St. Clair et al. (9) reported. Whenthis patient then received AZT for 2 mo, insensitivity to AZTincreased rapidly (Fig. 2 and Table 1). This increase wasaccompanied by a genotypic reversal to wild-type Leu-74 innearly 80% of the clones, whereas all the clones retained theThr-215 -* Tyr substitution (Table 1). This rapid loss of theLeu-74 -+ Val substitution within 2 mo of ddI discontinuation

Dow

nloa

ded

by g

uest

on

June

6, 2

020

Proc. Natl. Acad. Sci. USA 90 (1993)

VIRUSSTRAIN

30 40 50 60 70 80 90VKQWPLTEEK IKaLvEICTE MEKEGKISKi GPENPYNTPV FAIKKKDSTK WRKLVDFREL NkrTQDFWEV

ERS200pre I--------- ---------- ----R----- R--------- ---------- ---------- ----------

ERS200pot ---------- ---------- ---------- -------------------A- ---V------ ----------

ERS201pr* ---------- ---------- ---------- ---------- ---------- ---------- ----------

ERS201pot--- -VERS203pre ---------- ---------- L-E------- ---------- ---------- ---------- ----------

ERS203post ---------- --------A- L-E------- ---------- ---------- ---V------ ----------

ERS204pr* ----s----- ---------- ---------- ---------- ------N--R ---------- ----------

ERS204po-t---------- --------I- ---------- ---------- ---------- ---------- ------L---ERS205pr ----------- ---------- -----VT-- ---------I ---------R ---------- -------R--ERS205p .t

100 110 120 130 140 150 160QLGIPHPAGL KkKKSVTVLD VGDAYFSVPL DkdFRKYTAF TIPSiNNETP gIRYQYNVLP QGWKGSPAIF

e r

ERS200pr. --------R- -R-------- ---------- ---------- ---------- -T-------- ----------

ERS200postERS201pre ---------- ---------- ---------- ---------- ---------- -T-------- ----------

ERS201post ---------- ---------- ---------- ---------- ---------- -T-------- ----------

ERS203pr. ---M---S-- ---------- ---------- ---------- ---------- ---------- ----------

ERS203poSt-------S-- ---------- -------I-- ---------- ---------- -V-------- ----------

ERS2O4preERS204postERS205pre ---------- ---------- ---------- ---------- ---------- --------F- ----------

ERS20 5post

170 180 190 200 210 220 230QsSMTKILEP FrKQNPdIvI YQYMDDLYVG SDLEIgQHRt KIEELRqHLL rWGfTTPDKK HQKEPPFLWMc e e k

ERS200pre ---------- ------------------ -----.------ - ------ H--- ---------- ----------

ERS200pot ---------- ---------- ---------- -----------------H--- ---------- ----------

ERS201pre ----I----- ---------- ---------- ---------- ---------- ---------- ----------

ERS201post ----I----- ---------- C--------- ---------- ---------- ---------- ----------

ERS203pre ---------- ---------- ---------- ---------- ---------W ----Y----- ----------

ERS203po-t---------- ---------- ---V------ ---------- --------W ----Y----- -------H--ERS204pr, ---------- ---------- ---------- ---T------ ---------- ---------- ----------

ERS204po-t---------- ---------- ---------- ---T------ ---------- ----Y----- ----------

ERS205pr. -------N-D- -------L-- ---------- ------R--- ---------- ---------- ----------

ERS205pot ---------D- -------L-- ---------- ---------- ---------- ---------- ----------

FIG. 4. Deduced amino acid sequences of RT in HIV strains isolated from patients receiving ddl. Four and eight clones were sequencedfor pretherapy strains and posttherapy strains, respectively. The following substitutions were found: in strain ERS200p,,, V211(3/4), G45R(1/4),G51R(1/4), G"R(1/4), k102R(1/4), 1142T(1/4), and q2O7H(4/4); in strain ERS2OOp.,t, T69A(2/8), L74V(8/8), and q2°7H(7/8); in strain ERS201pe,1142T(4/4) and T165I(4/4); in strain ERS201POSt, L74V(8/8), P142T(8/8), T165I(8/8), and Y181C(7/8); in strain ERS203pre, M41L(4/4), K43E(4/4),I94M(1/4), A"S(4/4), L210W(4/4), and T21 Y(4/4); in strain ERS203p.t, T39A(6/8), M41L(8/8), K43E(3/8), L74V(8/8), A98S(8/8), V1181(4/8),P142V(4/8), M184V(5/8), L21OW(6/8), T215Y(7/8), and LmH(7/8); in strain ERS204pFC, P25S(1/4), D67N(1/4), K70R(3/4), and E194T(4/4); in strainERS204post, T39I(2/8), F17L(3/8), E'94T(8/8), and T215Y(8/8); in strain ERS205.r, 147V(1/4), S48T(4/4), V60I(1/4), K70R(1/4), W"R(1/4),L149F(1/4), I'67N(1/4), E16D(4/4), I178L(2/4), and Q197R(1/4); and in strain ERS205post, E169D(8/8) and I178L(8/8).

AMINO ACID SEQUENCE

contrasts strikingly with the persistence ofthe Thr-215 -- Tyrsubstitution in nearly 90% of clones despite 26 mo of ddItherapy (Table 1).

DISCUSSIONIn this study, three of four post-AZT/ddC strains showed nosignificant decrease in sensitivity to ddC, although all strainshad developed high levels ofAZT insensitivity. In a separatepediatric study, we have isolated an additional seven pairs ofpre- and post-AZT/ddC strains and have seen the develop-ment ofa high level ofAZT insensitivity in four such patients,despite the addition of ddC to the therapy (R.N.H., T.S.,P. A. Pizzo, and H.M., unpublished observations). Theseresults suggest that although the alternating regimen ofAZTand ddC may reduce the side effects of each drug andconceivably have clinical advantages over monotherapy, thecombined AZT/ddC therapy does not prevent the develop-ment of HIV-1 variants less susceptible to AZT.One AZT/ddC-insensitive strain, HIV-11o3post41mo, had no

previously reported AZT-related mutations, but seven dis-

tinctive amino acid substitutions were found in this strain.These results suggest either that the amino acid substitutionsin HIV-1l13post41mo provided a replicative advantage in thepresence of AZT that exceeded that provided by combina-tions of the previously reported AZT-related substitutions orthat the mutations observed in HIV-1lO3post-lmo caused con-formational changes incompatible with the previously re-ported AZT-related substitutions. The most remarkable sub-stitution identified in HIV-11o3pst is the substitution forglutamine at position 151 with methionine [Gln-151 -* Met(8/8)]. This residue is part of a highly conserved amino acidsequence (Leu-Pro-Gln-Gly, amino acids 149-152 in Fig. 3),termed motif B, which is seen in each of 26 different retro-viruses (13, 14) including HIV-1, HIV-2, simian immunode-ficiency virus, human T-cell leukemia virus, and a number ofanimal retroviruses. Most recent x-ray crystallography of theHIV-1 RT has suggested that Gln-151 lies in (or adjacent to)the 3-strand 8, which is located close to the putative poly-merase active site (15). Thus, the Gln-151 Met substitutioncould alter the interaction between the RT and dNTP orchange the interaction between the enzyme and the template.

Medical Sciences: Shirasaka et al. 565

Dow

nloa

ded

by g

uest

on

June

6, 2

020

566 Medical Sciences: Shirasaka et al.

Table 1. Phenotypic and genotypic changes of HIV-1 strainssequentially isolated from a patient receiving AZT or ddI

Drug sensitivity Clones ClonesAntiviral (CNmo), ALM examined Residue notherapy AZT ddC ddI no. 74 215 (%)

12-mo AZT 32 1 17 4 WT M 4(100)Additional26-mo ddI 2.3 3.2 22 8 M M 7 (88)

M WT 1(12)Additional2-mo AZT 32 3.2 22 9 WT M 7 (78)

M M 2 (22)

Residue 74 represents leucine and valine for wild-type (WT) andmutant (M) amino acids, respectively; whereas residue 215 repre-sents threonine and tyrosine for wild-type and mutant amino acids,respectively.

In the current study, no appreciable decrease in sensitivityto ddI was detected in five different post-ddl HIV-1 strains,although the decreased sensitivity could have been so modestin degree that the difference was not readily detected. St.Clair and coworkers (9) have reported that HIV-1 strainsisolated after ddI therapy developed decreased sensitivity (upto 26-fold) to ddI as compared with pretherapy isolates. It ispossible that their use of RT activity and different PBMCbatches rather than the use of biologically more stable p24Gag protein and a single PBMC batch for comparison couldhave resulted in greater variability.The acquisition of the Leu-74 -- Val substitution in addi-

tion to the Thr-215 -* Tyr substitution appears to provideHIV-1 with a replication advantage in the presence of ddI;however, in the absence of ddI, the Leu-74 -- Val substitu-tion may cause a significant replication disadvantage. Thisresult may explain why HIV-1 mutants with the Leu-74 -+

Val substitution were rapidly eliminated from the viruspopulation when therapy no longer included ddI (Table 1),although elucidation of the conformational and functionalbasis for the decreased susceptibility to AZT and the Leu-74-* Val substitution-related restoration ofAZT sensitivity willdepend on continued analysis ofRT by x-ray crystallographyand related technologies (15, 16).HIV-1 apparently undergoes selection pressure to develop

reduced susceptibility to AZT. RT of AZT-insensitive vari-ants could, in theory, discriminate between AZT triphos-phate and dTTP by the presence or absence ofthe tubular andbulky 3'-azido group (17). In this regard, the close structuralresemblance of ddATP and ddCTP (the putative active moi-eties of ddI and ddC, respectively) to their natural corre-sponding nucleotides (dATP and dCTP, respectively) maymake it more difficult for the virus to mutate to efficientlyexclude ddATP or ddCTP, while preserving the function ofRT. Indeed, molar-refraction data (18) indicate that an azidogroup is about three times larger than a hydroxyl group,whereas a hydrogen molecule is 2-fold smaller than ahydroxyl group. It is possible that the mutations affect the

template-binding site in RT and change the ability of theenzyme to discriminate between natural dNTP and ddNTP,as proposed by Kohlstaedt et al. (15). However, the rela-tionship between the structural resemblance of the nucleo-tides and development of drug resistance is as yet highlyhypothetical and requires future assessment.

We thank Robert Wittes, Marvin Reitz, John Driscoll, SamuelWilson, and John Erickson for helpful discussions and the medicaland nursing staffs of the Medicine Branch of the National CancerInstitute for their help.

1. Larder, B. A., Darby, G. & Richman, D. D. (1989) Science 243,1731-1734.

2. Larder, B. A. & Kemp, S. D. (1989) Science 246, 1155-1158.3. Kellam, P., Boucher, C. A. B. & Larder, B. A. (1992) Proc.

Nati. Acad. Sci. USA 89, 1934-1938.4. Mitsuya, H. & Broder, S. (1986) Proc. Nati. Acad. Sci. USA 83,

1911-1915.5. Mitsuya, H., Yarchoan, R. & Broder, S. (1990) Science 249,

1533-1544.6. Yarchoan, R., Perno, C.-F., Thomas, R. V., Klecker, R. W.,

Allain, J.-P., Wills, R. J., McAtee, N., Fischl, M. A., Dubin-sky, R., McNeely, M. C., Mitsuya, H., Pluda, J. M., Lawley,T. J., Leuther, M., Safai, B., Collins, J. M., Myers, C. E. &Broder, S. (1988) Lancet 1, 76-81.

7. Yarchoan, R., Mitsuya, H., Thomas, R. V., Pluda, J. M.,Hartman, N. R., Perno, C.-F., Marczyk, K. S., Allain, J.-P.,Johns, D. G. & Broder, S. (1989) Science 245, 412-415.

8. Lambert, J. S., Seidlin, M., Reichman, R. C., Plank, C. S.,Laverty, M., Morse, G. D., Knupp, C., McLaren, C., Petti-nelli, C., Valentine, F. T. & Dolin, R. (1990) N. Engl. J. Med.322, 1333-1340.

9. St. Clair, M. H., Martin, J. L., Tudor-Williams, G., Bach,M. C., Vavro, C. L., King, D. M., Kellam, P., Kemp, S. D. &Larder, B. A. (1991) Science 253, 1557-1559.

10. McConway, M. G., Chapman, R. S., Beastall, G. H., Brown,E., Tillman, J., Bonar, J. A., Hutchinson, A., Allison, T.,Finayson, J., Weston, R., Beckett, G. J., Carter, G. D., Car-lyle, E., Herbertson, R., Blundell, G., Edwards, W., Glen,A. C. A. & Reid, A. (1989) Clin. Chem. 35, 289-291.

11. Saiki, R. K., Scharf, S., Faloona, F., Mullis, K. B., Horn,G. T., Erlich, H. A. & Arnheim, N. (1985) Science 230, 1350-1354.

12. Fitzgibbon, J. E., Howell, R. E., Harberzettl, C. A., Sperber,S. J., Gocke, D. J. & Dubin, D. T. (1992) Antimicrob. AgentsChemother. 36, 153-157.

13. Xiong, Y. & Eickbush, T. H. (1990) EMBO J. 9, 3353-3362.14. Poch, O., Sauvaget, I., Delarue, M. & Tordo, N. (1989) EMBO

J. 8, 3867-3874.15. Kohlstaedt, L. A., Wang, J., Friedman, J. M., Rice, P. A. &

Steitz, T. A. (1992) Science 256, 1783-1790.16. Arnold, E., Jacob-Monila, A., Nanni, R. G., Williams, R. L.,

Lu, X., Ding, J., Clark, A. D., Jr., Zhang, A., Ferris, A. L.,Clark, P., Hizi, A. & Hughes, S. H. (1992) Nature (London)357, 85-89.

17. Cameron, A., Mastropaolo, D. & Cameron, N. (1987) Proc.Nati. Acad. Sci. USA 84, 8239-8242.

18. Hansch, C., Leo, A., Unger, S. H., Kim, K. H., Nikaitani, D.& Lien, E. J. (1973) J. Med. Chem. 16, 1207-1216.

19. Myers, G., Berzofsky, J. A., Korber, B. & Smith, R. F. (1991)Human Retroviruses and AIDS (Los Alamos National Lab.,Los Alamos, NM).

Proc. Nad. Acad Sci. USA 90 (1993)

Dow

nloa

ded

by g

uest

on

June

6, 2

020